Recurrence and progression in stage T1G3 bladder tumour with intravesical bacille Calmette-Guérin (Danish 1331 strain)



Objective  To report recurrence and progression rates in patients with T1G3 superficial bladder carcinoma treated with intravesical bacille Calmette-Guérin (BCG, Danish 1331 strain) after complete transurethral resection.

Patients and methods  Data from the records of 111 patients with T1G3 bladder carcinoma treated between January 1991 and December 1999 were analysed for recurrence, progression, salvage therapy and survival.

Results  Of the 111 patients with T1G3 bladder tumours, 69 had intravesical BCG therapy, 20 radical cystectomy and 22 only transurethral resection (TUR). Of the 69 patients receiving BCG therapy 37 (54%) had no recurrence, and 24 (35%) had a recurrence that was not muscle-invasive (Ta/T1) and were treated with TUR only. The remaining eight (12%) progressed to muscle invasion and had salvage cystectomy. During the follow-up six patients died, four from disease and three from other causes, while the remaining 63 are alive and well. Of the other 42 patients, 15 are alive after radical cystectomy and 18 after TUR.

Conclusion  This series further confirms the benefits of intravesical BCG (Danish 1331) in an adjuvant setting; furthermore, this treatment facilitates bladder preservation by reducing recurrences and delaying the progression in many patients.


Tumours of the bladder that are not muscle-invasive are subcategorized into those limited to mucosa (Ta) and those involving the lamina propria (T1). Endovesical chemoprophylaxis or BCG therapy is mandatory in T1 tumours, while it is optional in Ta tumours. A subset of T1 tumours, those designated T1G3, is associated with a worse outcome because of their aggressive biology [1] and hence require aggressive TURBT and endovesical adjuvant therapy as primary treatment [2]. Historically, early cystectomy for T1G3 tumours had yielded good results but currently the emphasis is on bladder preservation. Various authors have reported success and failure in the management of T1G3 tumours using various strains of intravesical BCG in an adjuvant setting [3–23]. We herein report our experience of managing 69 patients with T1G3 bladder tumours using endovesical BCG (Danish 1331 strain).

Patients and methods

From January 1991 to December 1999, 840 patients with bladder cancer that was not muscle-invasive were treated at the genitourinary service of our hospital. Of these, 111 (13.2%) had T1G3 tumours, i.e. tumour involving the lamina propria of the bladder wall, with grade III histology. The histopathology of all these patients was reviewed by a dedicated and experienced uro-oncopathologist. No attempt was made to identify carcinoma in situ (CIS) in these patients. After a complete evaluation all 111 patients underwent cystoscopy and TURBT. Sixty-nine of the patients received endovesical BCG therapy 2–3 weeks after TURBT, given in six sittings as weekly instillations of 120 mg BCG (Danish strain 1331, Guindy, Madras, India; each ampoule containing 40 mg). The criteria for discontinuing intravesical BCG therapy were high-grade fever, severe systemic symptoms, haematuria and cystitis, as reported previously [13,14]. None of the patients had maintenance therapy but two required a second course of BCG.

The follow-up comprised urine cytology and cystoscopy 3-monthly for the first 2 years and 6-monthly for the next 3 years. Recurrence or progression was managed by repeat TURBT or cystectomy, depending on the nature and invasiveness of the recurrence. Patients undergoing cystectomy were followed with basic blood investigations, ultrasonography, CT of the abdomen and pelvis, and a loopogram (optional) every 3–6 months. The median (range) follow-up was 45 (18–120) months.


Of the 111 patients with T1G3 tumours, 69 were treated with endovesical BCG; 20 of the remaining 42 patients underwent radical cystectomy, of whom 15 are alive and disease-free. Five patients had recurrent disease, of whom three died. Of the remaining 22 patients, six are alive and disease-free. Eight patients had a recurrence that was not muscle-invasive, managed by TURBT (one), TURBT with BCG (two) and radical cystectomy (five). Eight patients had muscle-invasive recurrences managed with cystectomy (six), chemotherapy (one) and one patient refused treatment. The 69 patients treated with intravesical BCG had a mean (range) age of 56.7 (28–80) years and a male : female ratio of 4 : 1. The data were analysed for time to first recurrence, time to first progression, treatment of recurrence and survival.

Of the 69 patients, 37 (54%) had no recurrence, 24 (35%) had other than a muscle-invasive recurrence (Ta, T1) and the remaining eight (12%) had recurrence with progression (T2). The median time to first recurrence was 18.1 months and to progression 12 months. Recurrences that were not muscle invasive in 24 patients were managed by endoscopic surgery in 22, and the histology of the recurrences was Ta in 12 and T1G1 in 10, while the remaining two had T1G3 and had an additional second course of BCG after TURBT. All eight patients with progression (T2) were treated with radical cystectomy (Table 1). During the follow-up, of the 69 patients, six died (four from disease and two from other causes) while the remaining 63 were alive and well at a median follow-up of 45 months.

Table 1.  The characteristics of eight patients who progressed after BCG treatment
Delay in radical cystectomy, monthsPathological
13–5 cm, carpet-like, two sites, 3–5n20pT2bN0M0
2< 3 cm, papillary, 2 sites, 2n18pT2aN0M0
3< 3 cm, papillary, 3 sites, CIS26pT1N0M0, with CIS
43–5 cm, not papillary, solitary15pT2aN0M0
53–5 cm, solid, solitary 6pT2aN0M0
63–5 cm, solid, solitary 9pT2aN0M0
73–5 cm, solid, solitary 4pT2bN0M0
83–5 cm, solid, solitary 7pT3aN0M0

Delayed cystectomy as salvage surgery was undertaken in eight of 69 patients (12%) for progression (Table 1). The cystectomy specimen showed T1 and CIS histology in one, T2 in six and T3 in one, while none had metastatic nodes. Seven of eight patients are alive while one with T3 histology died from distant metastasis after 12 months. Kaplan-Meier curves of recurrence-free and overall survival are shown in Fig. 1.

Figure 1.

Figure 1.

Kaplan Meier curves for a, recurrence-free survival and b, overall survival.

Figure 1.

Figure 1.

Kaplan Meier curves for a, recurrence-free survival and b, overall survival.


Although included in the non-muscle-invasive category, T1G3 bladder tumours have become a distinct entity because of their higher potential for recurrence and progression [1]. The aim of treatment in these patients is to retain the functioning bladder without compromising survival. Previously these patients were either treated by TURBT alone [1,24] or by radical cystectomy [15,24–26]; both treatments had benefits and morbidity. TURBT alone has the danger of unsalvageable relapses, leading to only a 50% 5-year survival [1,16,17] while radical cystectomy provides excellent survival (70–90% at 5 years) but has associated morbidity of a stoma or management of a neobladder. In contrast, delayed cystectomy after recurrence or progression gives a 5-year survival rate of 50–60%, according to the proponents of early cystectomy [18–21]. Hence, endovesical adjuvant therapy gives both advantages, i.e. control of disease and bladder preservation, to provide a good quality of life. The addition of endovesical chemoprophylaxis in an adjuvant setting does not add to survival [17] but Bono et al.[22] reported success with doxorubicin in T1G3 disease. The reason for the poor results of chemoprophylaxis was considered to be the lesser surface contact with suburothelial stroma, which is the deepest layer of a T1 tumour [24].

The emergence of intravesical BCG in the late 1970s, and subsequent encouraging reports of endovesical BCG in the adjuvant setting, supported its use as the standard treatment. In superficial tumours (Ta, T1) the net benefit of BCG therapy appears to be an increase in the recurrence-free interval and a delay in cystectomy. Among the different strains of BCG, Morales et al.[23] first administered BCG to humans using the Aarmand-Frapier strain. Subsequently, many other strains have been introduced, e.g. Pasteur, Tice, Connaught, Glaxo, Evans, Tokyo, Dutch (RIVM) and Moreau. In our institution [14] we use the Danish 1331 strain, which is as effective as the others (Table 2). The side-effects of this strain are also comparable with other strains [14].

Table 2.  Previous reports of BCG treatment for patients with stage T1 grade 3 tumours
StudyNo. of
 [4] 473612
 [3] 44271643Connaught & Tice
 [5] 86441959Pasteur
 [6] 24502522
 [7] 30340639Pasteur
 [8] 25400463
[20] 50281242Pasteur
[16] 62200046
[17] 50321252
Present 6934.711.745Danish 1331

The higher (13%) incidence of T1G3 tumours in our practice is largely caused by the referral pattern. BCG (Danish 1331) therapy gave good results in the present series in preventing recurrences in 54% of the patients and down-staging the recurrences in 35% (Ta,T1 but < G3 recurrences). Thus, the net benefit of BCG therapy is preservation of the bladder in 88% of patients. Similar results were reported by others with different strains of BCG [3–8]. The T1G3 recurrences after BCG were treated with either a second course of BCG in a few or radical cystectomy in most patients. Some consider that maintenance therapy delays recurrences and progression; we have not used maintenance therapy because of toxicity and compliance. Progression after BCG therapy in the present series appears to be comparable with that in other series and was treated by radical cystectomy, giving higher rates of bladder conservation (Table 2). Danish strains of BCG give comparable results to other strains for recurrence pattern, increased recurrence-free interval and delay in progression.

In conclusion, this series further confirms that intravesical BCG Danish 1331 is an effective adjuvant therapy for T1G3 tumours. Recurrences or progression can be safely managed by repeat TURBT or cystectomy. This results in an overall higher rate of bladder conservation and equal survival rates. Prospective randomized trials would be required to conclusively quantify these results.

J.N. Kulkarni, C-14 TMC Qtrs, Mandala, Sion Trombay Road, Mumbai – 400088, India.